February 19, 2013
The Holy Grail For Monitoring Your Portfolio
What's the most important source of strategic information for managing your investment portfolio? Is it the outlook for each of the markets that are targeted by your strategy? Estimates of risk? Expectations for interest rates or the business cycle? Actually, most of what you need to know comes directly from the portfolio--your portfolio. To be precise, the shifting allocation weights contain the lion's share of critical data you need for making informed investment decisions and generating a satisfactory risk-adjusted return through time.
This is obvious, or at least it should be. But in the rush of analysis and commentary on all things financial and economic in the world, this subtle but enormously crucial point is easily overlooked (or perhaps ignored altogether?). A huge mistake.
Consider a simple example. You have a portfolio that’s initially weighted 60% stocks and 40% bonds (S&P 500 and Barclays Aggregate, respectively). Those are your target allocations and you plan on rebalancing when those allocations move substantially away from their targets. What defines “substantially”? That’s a topic for another day. For now, let’s imagine that the equity weight has fallen to, say, 40%. Is it time to rebalance back to 60%? In other words, is it time to buy stocks? This is a rather large clue that the answer is “yes”... for this portfolio.
But shouldn’t you also analyze the outlook for the stock market, the economy, and run a battery of statistical and econometric tests to figure out if this is a good time to buy stocks based on the internal dynamics of the equity market? Maybe. Those type of things can be useful overlays. Maybe you temper your rebalancing, based on those additional inputs. Or maybe you ramp up the rebalancing and reweight stocks to exceed the target allocation. But you'd need one deeply convincing set of overlay data to convince you not to rebalance at all.
In short, history strongly suggests that you can do quite well by making decisions primarily if not exclusively from the fluctuating weights in your portfolio. One reason why this is true is that future returns for a given asset class tend to be negatively correlated with the current allocation changes of that asset class (relative to your target allocation). Intuition suggests as much. After all, if your equity allocation is substantially above your target weight, that’s probably because equities have rallied sharply, and vice versa. Buy low, sell high, as they say. Duh! This can be dangerous with individual securities, but it's far more compelling when we're talking the major asset classes and their primary subdivisions.
For instance, consider the graph below, which compares the unmanaged equity allocation (S&P 500) through time for a 60% stocks/40% bond portfolio. The gray area represents the equity allocation. It’s initially set at 60% at the end of December 1975, and left to wander unmanaged through the end of 2011. The dark red line represents the year-ahead return for the S&P 500.
The general relationship is that high returns in the year ahead tend to be linked with relatively low equity allocations in the here and now, and vice versa. The message in this simple 60/40 portfolio is that when your equity allocation falls (rises) relative to bonds, there’s a stronger case for buying (selling) equities via rebalancing.
To be sure, this relationship isn’t perfect, at least not every day. Nothing in finance works flawlessly. But the rebalancing observation above is a reasonable rule of thumb. It may be broken at times, or mildly modified, but in the grand scheme of portfolio strategy it's quite practical. Let’s also recognize that the analysis and review can get a bit more complicated if you have multiple asset classes. But you can sort out the valuable information from the market noise by routinely monitoring the portfolio weights in search of relatively strong rebalancing signals.
For instance, let’s take another example: an internationally diversified 60% stocks/40% bonds portfolio that was launched at the close of 2011 using four ETFs to represent each of the markets. The equity slice is evenly split between domestic and foreign stocks, and the same applies to the bond side. The portfolio is left unrebalanced for more than a year. Here’s how it looked as of last Friday, February 15, 2012. Note the perspective on how the current allocations stack up against the historical ranges:
Does the information in the chart above tell us that it’s a good time to rebalance? The answer depends on several factors, including the investor’s time horizon, risk tolerance, and perhaps his tax situation, to name a few items. In short, there’s no one right answer for everyone.
On the other hand, it's essential for everyone to routinely look at the type of information as shown in the chart above. I run various forms of this analysis frequently on my own portfolios, and I also generate comparable data for clients on a few consulting projects. Invariably, this information delivers a large piece of the strategic information that guides decisions on how to manage the asset allocation through time.
In fact, once you start looking at this type of information regularly, you tend to filter out a lot of the noise about markets that comes from the usual suspects. Why? Because most of what you need to know to manage your investments successfully comes from within the portfolio structure. But there’s a catch: you have to be looking, and you should be looking fairly frequently. That doesn’t mean that you’re rebalancing frequently. But quite a lot of the strongest rebalancing opportunities come and go quickly. You might miss those ripe moments if you’re only looking at the allocation data monthly or quarterly or (even worse) once a year.
Finally, don’t fool yourself into thinking that if you watch the markets directly you’ll have good intuition on when to rebalance your portfolio. That may be true, but sometimes it’s not. Why? Because the interaction of the asset classes within your portfolio may deliver signals that aren’t obvious when looking at the various markets directly and individually. Remember: your portfolio is unique, which implies that the rebalancing signals may not correlate closely with the trends in, say, the S&P 500.
The bottom line: monitor and analyze your allocation data regularly. Weekly is probably reasonable. Once a month or beyond is easier, of course, but you may be overlooking critical information—information that only your portfolio can provide.
Posted by jp at February 19, 2013 9:55 AM
In the long run, a simple strategy of automatically rebalancing once a year, for instance, is likely to do well vs. a broad array of opportunistic strategies that attempt to optimize the process. In the rebalancing chapter of my book (Dynamic Asset Allocation), I reviewed a number of studies on this question. The results are all over the map. The issue is that the best rebalancing strategy depends on a number of factors that can't be known in advance. As a result, there's a strong case for automatic rebalancing to average out the results, but minds will differ. The one point that most of the literature agrees on is that you should perform rebalancing periodically, somewhere between once a quarter up to once every two years. Of course, this is only makes sense if you're broadly diversified in the first place.
Posted by: JP at February 21, 2013 6:34 AM
You could potentially devise a simple rebalancing algorythm, backtest it, and form a robust rule that would take away the emotional aspect of it, as we know that robust models overwhelmingly outperform individual discretionary decisions. What do you think?
Great blog by the way.
Posted by: Bruno V. at February 19, 2013 11:00 PM